- Glif tests
- Trading card
Trading card
Every model got the same prompt. Here's what 26 models made, best first.
Results
Won
FLUX.2 Flex
Score 87 · 2 wins, 0 losses
#2 of 12
GPT Image 2.5
Score 84 · 3 wins, 0 losses · Judge 9/10
#3 of 12
Muse Image
Score 73 · 2 wins, 1 loss · Judge 9/10
From 14 blind votes by the Glif team and 4 AI judge scores. Scores are out of 100, and 50 is average for this test. See the full text-to-image leaderboard
How we score
Scores on this page come from this test alone and compare only the models that ran it. Leaderboard scores pool every test of the same kind.
Every model runs the same Glif tests. Team members see two outputs side by side without the model names and pick the one that better fits the test, or call it a tie. An AI judge also scores every output from 0 to 10.
A Bradley-Terry fit turns all those head-to-head results into one rating per model. Ties count as half a win, and each AI judge score counts as half a vote.
The score is the predicted chance people prefer this model's output to the average model's on the same board, so 50 is average. A board is one kind of task, like text-to-image. Scores only compare models on the same board.
The 90% range comes from resampling the votes and refitting many times. Overlapping ranges are close to a tie. A model ranks once it has 20 votes' worth of ratings on a board. Until then it shows no score, only how close it is.
What we look for
Produces a plausible trading-card layout with legible text. Anime style.
Prompt
“A full trading card of an anthropomorphic junkdealer hamster with an arm prothesis. Anime style artwork. Late 1990s japanese TCG card design. Text should feature stats, costs, moves, and flavor text in english.”
Tests for
Composition, typography, style